Water cooled solar panels for significant output boost

[posplayr]

I just read most of this thread and found it very interesting for a variety of reasons.

It is obvious that there are some very intelligent people here.

PS I introduced myself here.

http://www.solarpaneltalk.com/showthread.php?21847-Hello-from-Tucson-PV-newbie&p=161489#post161489

[LordAC]

Hi!

I'm doing this experiment in my thesis.
Basically this will have a closed loop water system. I'll see if what I get in the solar panels efficiency will somehow compensate for the use of water and electricity (water pump).
In addition to this, the idea is also clear the panel's dirt and dust (this greatly decreases the efficiency).

I believe the main problem here the water evaporation and the water "contamination".

[J.P.M.]

The method of cooling used here is hard to understand. By running liquid water over top of the panels, some amount of sunlight will be reflected or adsorbed by the water, reducing the transmitted irradiance.

I would disagree with that statement. I have experimented with this principle during maintenance and module cleanings. Water does not block irradiance, it amplifies it, actually the water would act more like a million little fresnel lenses in its effect on the module. I watch production variances during regular cleanings. In a few experiments we have just misted the modules to see what happens and the production spikes upward until the mist evaporates. Now when I say mist, this is not enough to cool a module down I am referring to a light mist on a full sunny day. We have tried this before and after cleaning and seen the same result. I would agree with the OP. And also with the others about water usage. but a once a year cleaning of an array could very well use more water than a system like this, that may also help the modules to stay clean in the process. I would say this data looks promising.

To clear some stuff up:

1.) The method of cooling here - water run over the top surface of a panel - is mostly a combination or mix of convective heat transfer and evaporative heat transfer, one mechanism more or less responsible than the other, or about equally, depending on the amount of water involved and the environmental conditions present at any one time.

2.) In theory and practice, and contrary to what you might believe, or have heard, or perhaps misunderstood, in this application, H2O does reduce the amount of irradiance reaching the outer glass surface, but not much and probably to the point of being immeasurably small for such applications, the effect being mostly masked by the effects caused by cooling, which are probably at least a couple orders of magnitude greater. As a 1st approximation, the reflection from an H2O surface is approx. the same as from glass, so that's not an issue. But, to say that water amplifies or enhances the amount of irradiance reaching a panel is simply not true. Water, is isotropic (non directional) in most of its properties, at least the optical ones, and most all of the rest, and so does not exhibit any properties of Fresnel lenses which are nothing if not directional. The directional part is what is necessary for increasing radiant flux over an area. The moving areas of enhanced light observable on the bottom of a pond that has a disturbed surface are from the combined effects of reflection from the uneven top H2O surface and Fresnel effects. The enhanced irradiance at those sites comes at the expense of reduced light levels over the rest of the bottom surface. Note that the effects are isotropic, at least over a mostly 2 dimensional surface like the bottom of a pond.

3.) The production spikes you note in a PV device are caused by reduced cell temperatures. The reduced cell temp. is due, for the most part, from the evaporative effects of the water, or the convective heat transfer if the evaporation is not sufficient to evaporate all the H2O. Such a scheme will work best on windy days and/or when the dew point is low(er). The increased panel output is solely due to lower cell temperatures. In principle, and to reiterate, the reduced irradiance at the outer glazing surface that the H2O film causes will reduce the output, but the reduction is so small as to be masked by the improvement that reduced operating temperatures bring.

4.) Not something I'd recommend as more than a thought experiment, but if the water was warm, or hot, say, much hotter than the panel, the effect of spraying it on an operating panel would be to increase the cell operating temp., thereby lowering the cell (panel) output, particularly to the degree that the evaporative effects could not overcome the temp. diff. between the hot(ter) water and the relatively cooler panel. This whole business is a problem in heat transfer.

5.) Lots of folks have tried to increase panel efficiency and thus output, and also thereby use the waste heat for other purposes (pool heating for example). The idea has been around for centuries. To date, such methods have not be been shown to be safe, practical, economical or workable for residential size systems. In general, at this time, such schemes are mostly the purview of DIYer's who have little/any idea of how to approach the situation. Amen and have fun, but without some grasp of the basics, most of such efforts are a wheel spin and have all been investigated long ago. But, hope springs eternal, just waiting for some new technology to make it all safe and practical.

[ncs55]

Thanks all I have learned a lot from this thread. J.P.M., What you say makes sense. I do know that the cooling effect was mostly what caused the spike. I however do not fully understand why water would block the irradiance. I was speculating and relating this to what I see say in my garden when I water the plants in sunlight and some of the water gets on the leaves. The droplets of water on the leaves in full sunlight actually burn the leaves in the exact footprint of the droplet. I came to the conclusion that the said droplet was actually magnifying the irradiance say like a fresnel or magnifying glass. After reading through this thread again and what you posted. I think maybe it is the factor of concentrated heat not irradiance that caused that effect. ?? I think personally that the heat needs to be extracted from the bottom of the module as was already posted, but there still needs to be a certain amount of airflow to let the rest of the heat not captured from below a path of least resistance to dissipate away from the module. We have a few modules that we are testing by actually opening up airflow channels in the top and bottom of the frames to see if the heat can naturally convect away from the modules and how much gain can be achieved by a simple module frame modification that the manufacturer engineers can implement in their designs. it is a long shot, but if by changing the frame design can achieve greater performance this way it would be a cost effective way of letting more heat escape and harvesting more power. It would also allow mostly for better performance in installs that are really close to the roof, in theory at least. I am not sure how many people have approached this problem from this perspective. I would welcome any data and any points of view on this theory.

[Technologist]

Couple of issues of simple physics.

1. The energy used to cool the panels is more than the extra power produced so you have net negative gain.
2. I understand why you are using a water softener to prevent scale build up, but the trade off anyone that lives near a sea is salt water is hell on panels. You are trading Calcium and Magnesium for Sodium Chloride (salt).

Unfortunately you are not the first to go down this road. Many have gone down this road before you and discovered the gain does not offset the energy used to run the pumps. That is why you don't see any systems for sale. Then you got the environmental whackos would eat you alive because you are wasting considerable amounts of water, especially in Perth where you have extremely low humidity means most of the water evaporates before you can collect it and recycle.

I actually did this 25-30 years ago when I was cooling my solar concentrating panels and when I built a solar powered "closed loop” stream turbine electrical generating station. The entire system ran in a partial vacuum and boiled water at just over room temperatures to generate electricity from the sun.

It takes 8,092 BTUs to evaporate one gallon of water or 2.372 kWh. With the little energy needed to raise the water to that level It would be easy to get a net energy gain by evaporating water on the panels but all the other problems mentioned are real.

[J.P.M.]

Thanks all I have learned a lot from this thread. J.P.M., What you say makes sense. I do know that the cooling effect was mostly what caused the spike. I however do not fully understand why water would block the irradiance. I was speculating and relating this to what I see say in my garden when I water the plants in sunlight and some of the water gets on the leaves. The droplets of water on the leaves in full sunlight actually burn the leaves in the exact footprint of the droplet. I came to the conclusion that the said droplet was actually magnifying the irradiance say like a fresnel or magnifying glass. After reading through this thread again and what you posted. I think maybe it is the factor of concentrated heat not irradiance that caused that effect. ?? I think personally that the heat needs to be extracted from the bottom of the module as was already posted, but there still needs to be a certain amount of airflow to let the rest of the heat not captured from below a path of least resistance to dissipate away from the module. We have a few modules that we are testing by actually opening up airflow channels in the top and bottom of the frames to see if the heat can naturally convect away from the modules and how much gain can be achieved by a simple module frame modification that the manufacturer engineers can implement in their designs. it is a long shot, but if by changing the frame design can achieve greater performance this way it would be a cost effective way of letting more heat escape and harvesting more power. It would also allow mostly for better performance in installs that are really close to the roof, in theory at least. I am not sure how many people have approached this problem from this perspective. I would welcome any data and any points of view on this theory.

The short answer to your first statement: It's because water absorbs solar radiation.

Water or any translucent or semi transparent material, be it solid or fluid is never 100 % transparent to solar radiation. Water, just like air, reflects and absorbs solar radiation. Taking air for a minute, if it did not reflect and absorb solar radiation, the normal (in the context of perpendicular to the direction of propagation) intensity would be the same at sea level as above the atmosphere - currently defined as 1,367 W/m^, +/- ~ 3% to account for the earth's elliptical orbit. Also, the irradiance increases seen at higher elevations above sea level would be absent.. The same effect can be seen in water. At some ocean depth, the environment is dark. That's because water reflects solar radiation at the surface and absorbs solar radiation as the sun traverses the water. At some depth all the solar radiation has been absorbed and it's lights out. How much decrease is a function of the path length through the water, but the point is, the radiation does not increase.

The medium through which the solar radiation passes absorbs some of the solar radiation. That's the mechanism responsible for the heating the medium. If solar radiation were not absorbed, water sitting in the sun would not heat up.Entropy at work for you and me.

I'm not a botanist, but I'd guess the discoloration you note may have something to do with chemistry or photosynthesis or some other mechanism. Now, it may be that water droplets, because of the roundness induced by surface tension could act as a lens of sorts. The light hitting a tiny part of the drop surface has it's angle for refraction changed by the water. Because the drop is (hemi)spherical, the surface geometry (roundness) will tend to focus the light to some degree on a small(er) portion of the surface under the drop, the same way surface ripples can be seen on the bottom of a swimming pool, and for the same reasons, and also similar to what allows something called a Campbell Stokes sunshine recorder to work (see the net). However, the total surface under the drop will "see" less overall sunlight. The (small) area of enhancement that's brighter will benefit at the expense of the rest of the area under the drop, which will see less light, with the overall effect being less total sunlight hitting the surface under the drop as entropy dictates and a heat (energy balance) will confirm, by knowing that the lost surface irradiance will show up as increased droplet temp. That's the simple version. It does not consider evaporative cooling effects - some other time, OK ?

As for cooling solar devices. The way to determine how energy is lost form a solar device, be it solar thermal or PV, or anything else is to do an energy balance on it.

Take a solar panel, sitting in the sun and happily producing electricity for its owner. Say the normal (again perpendicular) irradiance is 1,000 Watts/m^2 , it's solar noon and the panel is perpendicular to the sun's direct rays. Also assume that for about 10 min. on either side of solar noon, the rest of the environmental conditions don't change enough to warrant worrying about them - clear skies, constant air temp. and a steady 5 m/sec. wind blowing, say from the west, the panel temp. is constant as is electricity flow. Everything is in what's called steady state and constant - an idealization, but sometimes real world conditions get close either by luck or by lab environment.

Now, imagine a "surface" around the panel, top, bottom and sides, much like the concept of a surface from solid geometry, and do some energy accounting. Because things are steady state, the energy that crosses the boundary in one direction - the solar going in - must equal the energy crossing the boundary in the other direction (going out), either as electricity or as heat for this simple example. The waste heat must, can and will be eliminated as dictated by the energy balance.

Say the panel is 1 M^2 and generates 175 Watts at STC. And say you've contracted me (having done such things several hundred time) to measure your panel temp. at the above conditions and I come up with 60 deg. C as the panel temp. Because of temp. effects the panel eff. has dropped from 17.5% to (.175)* (1- 0.005(60-25) = .144 = 14.4%. So, 1,000-144 = 856 Watts will be rejected one way or the other - that's what a heat balance dictates. The natural cooling methods - forced convective heat transfer via the wind or natural convection via gravity, or thermal radiation heat transfer, from the top, bottom and sides of the panel to the surroundings, and some small amount of conduction heat transfer through the racking. All those paths must add up to 856 Watts of heat transfer.

The amount of heat transferred = The sum of all: (the rate of heat transfer for each mechanism) * ( temperature difference each mechanism "sees").

Since the temp. of the panel is the same for each mechanism, the more heat transfer mechanisms there are, AND/OR the more efficient the heat transfer mechanisms are at removing heat per deg. of temp. diff., the cooler the panel will run.

A gross oversimplification to make a point: Say all the heat transfer is from wind blowing across the top and bottom of the array (only) and the air temp = 30 C. That means that each degree of temp. diff. removes heat from the panel at a rate of (856)/(60-30) = 28.5 Watts. Now change (drop) the air temp. to 20 C. What that means is that while the rate of 28.5 Watts/m^2 per deg. C. stays about the same, panel temp. must drop because it can only reject 856 Watts (for this simple example). So the panel temp. will drop to (856)/ 28.5 = (panel temp. - 20) == -->> panel temp. = 50.0 C. (Which is also an example of why panels run cooler when it's cold.)

Now, apply some super doper closed loop cooling device or modification across the back of the panel that uses water as the cooling medium at, say, 20 deg. C. that will produce a heat transfer rate across the back of the panel of 200 W/m^2 per deg. C of temp. diff., water to panel, and forget about the front side convection loss for now. Since 856 Watts must will be removed, the temp. diff. between the panel and the cooling water will be: (856)/200 = 4.3 C, giving a panel temp. of 20+ 4.3 = 24.3 C. Kind of cool huh ?

Because of the added cooling at the back, and now going back to the top side convection loss, a simplified and not strictly correct way to do it is to say the front side loss per degree of temp. diff., panel to air, will be about half the front + back loss, or 28.5/2 = 14.2 W/m^2 per deg. C. So, (and using the lower 20 C. air temp.) the front side heat transfer ~ 14.3*( (24.3-20) = ~61.5 Watts.

What will happen in this case is that the panel temp. will float somewhere between 24.3 C. and a bit more or less but always be such that between the heat loss rates and temp. differences of all the heat transfer mechanisms running in parallel, and all the electricity produced, they will all fluctuate very slightly minute to minute (or second to second) in such a way as to always, between all energy transfer mechanisms, equal the 1,000 Watt input all for the purpose of maintaining that energy balance.

Now, the backside aux. cooling, as shown above, can be a very effective way to cool things. My numbers were cherry picked to be easy to follow, but they are not too far off reality. Forced convection heat transfer coefficients using water can easily be 1 or 2 orders of magnitude more than coefficients using air. The pumping costs are low compared to gas heat transfer, but the construction and maint. costs are higher. Overall, its a bunch of tradeoffs, but water cooling takes the price in terms of effective shock value cooling ability.

However, note that while the panel temp. dropped quite a bit, and the efficiency improved to essentially STC eff. (with an ~ 24.3 C. operating temp.), the added output amounted to about 21 Watts or 12.5% or so. A cost analysis might show that aux. forced cooling might be effective for large arrays, but I suspect the level and type of design necessary to achieve the #'s I've shown with either water or air would be rather costly and impractical for industrial or municipal size plants, and out of the question for residential type large installs. I cite the dearth of such schemes at this time as support for that statement.

Lastly (Whew !), I'd suggest that modifications to the racking or the backside of panels or similar methods probably won't do much to increase the heat transfer from the backside of an array. First off, any meaningful flow is mostly forced convection wind driven. Second, the back (or underside) flow is more likely turbulent or transition zone (laminar to turbulent) flow, with most of the turbulence being due to the obstructions - the racking, posts or supports. That's good for heat transfer because turbulence is a better flow regime for heat transfer than laminar or streamline. And so, "smoothing out" or streamlining the array structure may do little good since turbulent flow is usually better for heat transfer. In any case, I'd not put a whole lot of faith in reliance on enhancements to natural (gravity driven) convection, which is usually so weak that it's easily overwhelmed by a even a slight breeze. I suppose such attempted enhancements can't hurt, but they're usually pretty unreliable and pretty costly for the benefit realized. It's been my experience that such benefits and their usual paucity are not well understood by those who try to make such things work. Been there, done that.

As usual, take what you want. Scrap the rest.

[shocksofmighty]

I built something similar to bouncingtigger. Going into it I wasn't expecting any massive gains, it really was just for fun. Taking into account all pump usage, I'm getting a 4% (daily) boost. This is judged by having two panels (I have two strings of panels on different levels) as controls that don't see water during the cooling. My per-panel resolution limit is only 5 minutes, but this is enough to show clearly that during the cooling phase (when the average of 4 panels that have temp. probes attached to the underside is over 40 C and AC production is over 4 kW) I get an average of ~11 W more (DC) per panel (32 panels not including controls ) that is being water cooled. The cooling period is generally ~370 minutes on hot days this month. This is in line with what the published literature on the subject says. I think that one real challenge is that it is much harder to scale any kind of water cooling system for larger arrays. This made it very impractical to have some kind of microsprinkler per panel, so I just use a gear-drive pop-up sprinkler on each of my strings. This makes it a bit harder to get even distribution over the panels, so there some panels that have 20-25 W boost over control and some with just 3-4. This could be overcome some by just continuous spraying, but I'm using rain water and my recollection is less than adequate at this point so I've empirically determined that 5.5 minutes on and 4 minutes off provides something close to a sweet spot right now.

I suppose my point of sharing is to counter the haters (sunking, russ) contending that there can be no net gains from the system. I do agree that it is totally impractical from an economic stand point, but that it can be a 'fun' engineering task. I mostly set out to do this to add another use for my Raspberry Pi and to learn some python!

[J.P.M.]

In most of commercial human endeavor, practical and cost effective trumps the laws of nature most every time. Economizer methods for small arrays are neither practical or cost effective at this time.

[ButchDeal]

I suppose my point of sharing is to counter the haters (sunking, russ) contending that there can be no net gains from the system.
I do agree that it is totally impractical from an economic stand point!

ok so you want to counter their point that there no net gains AND you agree that it is impractical from an economic stand point.
seems like you are not doing so well at countering their point.

[shocksofmighty]

ok so you want to counter their point that there no net gains AND you agree that it is impractical from an economic stand point.
seems like you are not doing so well at countering their point.

You're conflating the two issues.
1. There is a net power gain from cooling the panels. i.e. You get more out than you put in from running the pump.
2. The difficulty in setting the system up doesn't make this viable economically from a production sense.

[ButchDeal]

You're conflating the two issues.
1. There is a net power gain from cooling the panels. i.e. You get more out than you put in from running the pump.
2. The difficulty in setting the system up doesn't make this viable economically from a production sense.

But that is exactly what they were saying. it might be possible to get a gain but the complexity and cost is not worth value of the gain. and problems.

[shocksofmighty]

But that is exactly what they were saying. it might be possible to get a gain but the complexity and cost is not worth value of the gain. and problems.

No, that's not what Sunking and russ were saying. I won't bother quoting the entire "laws of physics" business, but just go back to post #2:

1. The energy used to cool the panels is more than the extra power produced so you have net negative gain.

This is false, and I was simply providing additional evidence towards demonstrating that.

Others (JPM, etc.) said that you might get a gain but that it doesn't make much sense from the economic side (certainly not from a production perspective). I was simply providing additional observations that they were right.

[Sunking]

No, that's not what Sunking and russ were saying. I won't bother quoting the entire "laws of physics" business, but just go back to post #2:


This is false, and I was simply providing additional evidence towards demonstrating that.

Others (JPM, etc.) said that you might get a gain but that it doesn't make much sense from the economic side (certainly not from a production perspective). I was simply providing additional observations that they were right.

Not False, just very simple high school physics of Thermodynamics and Conservation of Energy.

[shocksofmighty]

Not False, just very simple high school physics of Thermodynamics and Conservation of Energy.

Can you let us know where you went to high school so I make sure I don't send my kids there?

There are some problems with this study, but have a look at Figure 7.

The net output energy from the PV panels is calculated based on subtracting the energy input from the energy output, where the energy input is the electrical energy needed for running the water pump during the cooling period. The net energy output from the PV panel as a function of the MAT is depicted in Fig. 7.

[J.P.M.]

Practical and cost effective trumps the laws of nature when human endeavor is concerned. Economizer schemes such as these, while possible, are neither practical or cost effective at this time.

[ButchDeal]

and certainly do not lead to "SIGNIFICANT output boost"
For the cost to do the install and maintain, you could more cheaply add more modules

[ncs55]

The short answer to your first statement: It's because water absorbs solar radiation.

Water or any translucent or semi transparent material, be it solid or fluid is never 100 % transparent to solar radiation. Water, just like air, reflects and absorbs solar radiation. Taking air for a minute, if it did not reflect and absorb solar radiation, the normal (in the context of perpendicular to the direction of propagation) intensity would be the same at sea level as above the atmosphere - currently defined as 1,367 W/m^, +/- ~ 3% to account for the earth's elliptical orbit. Also, the irradiance increases seen at higher elevations above sea level would be absent.. The same effect can be seen in water. At some ocean depth, the environment is dark. That's because water reflects solar radiation at the surface and absorbs solar radiation as the sun traverses the water. At some depth all the solar radiation has been absorbed and it's lights out. How much decrease is a function of the path length through the water, but the point is, the radiation does not increase.

The medium through which the solar radiation passes absorbs some of the solar radiation. That's the mechanism responsible for the heating the medium. If solar radiation were not absorbed, water sitting in the sun would not heat up.Entropy at work for you and me.

I'm not a botanist, but I'd guess the discoloration you note may have something to do with chemistry or photosynthesis or some other mechanism. Now, it may be that water droplets, because of the roundness induced by surface tension could act as a lens of sorts. The light hitting a tiny part of the drop surface has it's angle for refraction changed by the water. Because the drop is (hemi)spherical, the surface geometry (roundness) will tend to focus the light to some degree on a small(er) portion of the surface under the drop, the same way surface ripples can be seen on the bottom of a swimming pool, and for the same reasons, and also similar to what allows something called a Campbell Stokes sunshine recorder to work (see the net). However, the total surface under the drop will "see" less overall sunlight. The (small) area of enhancement that's brighter will benefit at the expense of the rest of the area under the drop, which will see less light, with the overall effect being less total sunlight hitting the surface under the drop as entropy dictates and a heat (energy balance) will confirm, by knowing that the lost surface irradiance will show up as increased droplet temp. That's the simple version. It does not consider evaporative cooling effects - some other time, OK ?

As for cooling solar devices. The way to determine how energy is lost form a solar device, be it solar thermal or PV, or anything else is to do an energy balance on it.

Take a solar panel, sitting in the sun and happily producing electricity for its owner. Say the normal (again perpendicular) irradiance is 1,000 Watts/m^2 , it's solar noon and the panel is perpendicular to the sun's direct rays. Also assume that for about 10 min. on either side of solar noon, the rest of the environmental conditions don't change enough to warrant worrying about them - clear skies, constant air temp. and a steady 5 m/sec. wind blowing, say from the west, the panel temp. is constant as is electricity flow. Everything is in what's called steady state and constant - an idealization, but sometimes real world conditions get close either by luck or by lab environment.

Now, imagine a "surface" around the panel, top, bottom and sides, much like the concept of a surface from solid geometry, and do some energy accounting. Because things are steady state, the energy that crosses the boundary in one direction - the solar going in - must equal the energy crossing the boundary in the other direction (going out), either as electricity or as heat for this simple example. The waste heat must, can and will be eliminated as dictated by the energy balance.

Say the panel is 1 M^2 and generates 175 Watts at STC. And say you've contracted me (having done such things several hundred time) to measure your panel temp. at the above conditions and I come up with 60 deg. C as the panel temp. Because of temp. effects the panel eff. has dropped from 17.5% to (.175)* (1- 0.005(60-25) = .144 = 14.4%. So, 1,000-144 = 856 Watts will be rejected one way or the other - that's what a heat balance dictates. The natural cooling methods - forced convective heat transfer via the wind or natural convection via gravity, or thermal radiation heat transfer, from the top, bottom and sides of the panel to the surroundings, and some small amount of conduction heat transfer through the racking. All those paths must add up to 856 Watts of heat transfer.

The amount of heat transferred = The sum of all: (the rate of heat transfer for each mechanism) * ( temperature difference each mechanism "sees").

Since the temp. of the panel is the same for each mechanism, the more heat transfer mechanisms there are, AND/OR the more efficient the heat transfer mechanisms are at removing heat per deg. of temp. diff., the cooler the panel will run.

A gross oversimplification to make a point: Say all the heat transfer is from wind blowing across the top and bottom of the array (only) and the air temp = 30 C. That means that each degree of temp. diff. removes heat from the panel at a rate of (856)/(60-30) = 28.5 Watts. Now change (drop) the air temp. to 20 C. What that means is that while the rate of 28.5 Watts/m^2 per deg. C. stays about the same, panel temp. must drop because it can only reject 856 Watts (for this simple example). So the panel temp. will drop to (856)/ 28.5 = (panel temp. - 20) == -->> panel temp. = 50.0 C. (Which is also an example of why panels run cooler when it's cold.)

Now, apply some super doper closed loop cooling device or modification across the back of the panel that uses water as the cooling medium at, say, 20 deg. C. that will produce a heat transfer rate across the back of the panel of 200 W/m^2 per deg. C of temp. diff., water to panel, and forget about the front side convection loss for now. Since 856 Watts must will be removed, the temp. diff. between the panel and the cooling water will be: (856)/200 = 4.3 C, giving a panel temp. of 20+ 4.3 = 24.3 C. Kind of cool huh ?

Because of the added cooling at the back, and now going back to the top side convection loss, a simplified and not strictly correct way to do it is to say the front side loss per degree of temp. diff., panel to air, will be about half the front + back loss, or 28.5/2 = 14.2 W/m^2 per deg. C. So, (and using the lower 20 C. air temp.) the front side heat transfer ~ 14.3*( (24.3-20) = ~61.5 Watts.

What will happen in this case is that the panel temp. will float somewhere between 24.3 C. and a bit more or less but always be such that between the heat loss rates and temp. differences of all the heat transfer mechanisms running in parallel, and all the electricity produced, they will all fluctuate very slightly minute to minute (or second to second) in such a way as to always, between all energy transfer mechanisms, equal the 1,000 Watt input all for the purpose of maintaining that energy balance.

Now, the backside aux. cooling, as shown above, can be a very effective way to cool things. My numbers were cherry picked to be easy to follow, but they are not too far off reality. Forced convection heat transfer coefficients using water can easily be 1 or 2 orders of magnitude more than coefficients using air. The pumping costs are low compared to gas heat transfer, but the construction and maint. costs are higher. Overall, its a bunch of tradeoffs, but water cooling takes the price in terms of effective shock value cooling ability.

However, note that while the panel temp. dropped quite a bit, and the efficiency improved to essentially STC eff. (with an ~ 24.3 C. operating temp.), the added output amounted to about 21 Watts or 12.5% or so. A cost analysis might show that aux. forced cooling might be effective for large arrays, but I suspect the level and type of design necessary to achieve the #'s I've shown with either water or air would be rather costly and impractical for industrial or municipal size plants, and out of the question for residential type large installs. I cite the dearth of such schemes at this time as support for that statement.

Lastly (Whew !), I'd suggest that modifications to the racking or the backside of panels or similar methods probably won't do much to increase the heat transfer from the backside of an array. First off, any meaningful flow is mostly forced convection wind driven. Second, the back (or underside) flow is more likely turbulent or transition zone (laminar to turbulent) flow, with most of the turbulence being due to the obstructions - the racking, posts or supports. That's good for heat transfer because turbulence is a better flow regime for heat transfer than laminar or streamline. And so, "smoothing out" or streamlining the array structure may do little good since turbulent flow is usually better for heat transfer. In any case, I'd not put a whole lot of faith in reliance on enhancements to natural (gravity driven) convection, which is usually so weak that it's easily overwhelmed by a even a slight breeze. I suppose such attempted enhancements can't hurt, but they're usually pretty unreliable and pretty costly for the benefit realized. It's been my experience that such benefits and their usual paucity are not well understood by those who try to make such things work. Been there, done that.

As usual, take what you want. Scrap the rest.

As far as modifying the module frame, that would be inexpensive, and would make for a lighter weight module. And the modified module that I am experimenting with does run cooler and has a better production value than the un modified module. Natural upward convection would be the most cost effective way to cool as opposed to all of your other statements about forced cooling etc. The idea here is to let as much heat out from under the module without using anything that draws more electricity to operate it. As far as any wind under the array, that would only help to cool the modules. If you have been here and done this can you share your data with me? It would be helpful in my experiment.

[SunEagle]

As far as modifying the module frame, that would be inexpensive, and would make for a lighter weight module. And the modified module that I am experimenting with does run cooler and has a better production value than the un modified module. Natural upward convection would be the most cost effective way to cool as opposed to all of your other statements about forced cooling etc. The idea here is to let as much heat out from under the module without using anything that draws more electricity to operate it. As far as any wind under the array, that would only help to cool the modules. If you have been here and done this can you share your data with me? It would be helpful in my experiment.

Certainly any type of heat "radiator" on or part of the panel should help reduce losses from high temps and show some type of efficiency improvement.

The problem is how much does it cost to manufacture that "radiator" as compared to how much energy you gain due to lower temps at the cells.

There is always a balance between making a product the performs the best it can yet still keeping the costs down to meet the net profit goal for the company.

Hopefully you may find a solution to getting rid of the heat on a solar panel without increasing the cost to make it.